AMD's Advancing AI 2026 summit in San Francisco featured a panel with Chris Lattner, Ramin Hasani, and Hassan Akbari. Lattner, creator of LLVM, called AI 'mid,' arguing that the current AI surface lay…
AMD's Advancing AI 2026 event highlighted a strategy to bypass NVIDIA's CUDA moat through architecture, as argued by Chris Lattner, creator of LLVM, who also claimed AI is "mid" to emphasize that the …
AMD's upcoming MI400-series MI455X accelerator, codenamed GFX1250 and targeting machine learning, will operate exclusively in Wave32 mode and support WMMA instructions, according to LLVM commits analy…
Chapel 2.9 shipped on June 18 with LLVM 22 and CUDA 13 support, a prototypical dynamic library loading feature for distributed parallel code, a browsable Mason package registry, and union types that h…
A new open-source compiler called uplpgsql compiles PL/pgSQL to native code, achieving 2-4x speedups on ordinary stored procedures and up to 22x on heavy procedural and numerical work, according to in…
LFortran and Enzyme enable exact gradients from legacy Fortran code without rewriting, turning validated physics solvers into differentiable layers for JAX. The experimental stack compiles Fortran to …
FreeBSD 16 development tree has removed the last GPL-licensed component, GNU dialog, making the base system entirely GPL-free for the first time. The cleanup, which took over a decade, simplifies comp…
A cubin (CUDA binary) is a standard ELF64 file with NVIDIA-specific sections that encode everything needed to load and launch a kernel, including machine code, parameter layout, and register allocatio…
LFortran, LLVM, and Enzyme have been combined to automatically differentiate legacy Fortran simulation code, producing exact gradients through a multi-step time loop that match an analytic answer. The…
A study of the LLVM open source project reveals how AI technologies are negotiated within developer communities, highlighting tensions between innovation and community governance. The research examine…
Intel has archived its Quantum Intrinsics open-source project, a C++ front-end for the Intel Quantum Compiler, along with other AI-related projects like OAP-MLLIB, as part of ongoing restructuring. Th…
A developer revived his custom bytecode virtual machine UVM and built a Clang backend to compile real software, successfully porting Doom to run on it at 27 FPS on a MacBook Air M5.…
A developer built Phobos, a tiny scale-free kernel language that compiles to PTX and runs on NVIDIA GPUs, achieving 76% of cuBLAS SGEMM performance on a 2080 SUPER. The language supports tile-DAG desi…
Chris Lattner created MLIR (Multi-Level Intermediate Representation) in 2018 at Google to solve the problem of fragmented compiler infrastructure across different hardware targets and programming mode…
NVIDIA's CUDA compiler pipeline transforms a simple vector addition kernel from PTX virtual assembly to SASS machine code through multiple compilation stages, including LLVM-based cicc and ptxas, befo…
LLVM merged three optimizations to its BumpPtrAllocator, reducing overhead in the fast path by skipping unnecessary realignment, removing a null check via a sentinel end pointer, and eliminating per-a…
A developer created Nyra, a programming language that combines Rust's memory safety with JavaScript's simplicity. Nyra features Rust-like ownership, JS-inspired syntax, optional typing, and an LLVM ba…
Researchers have published a paper arguing that the design principles underlying MLIR, a compiler infrastructure for AI models, have formal theoretical correspondences that can clarify abstraction com…
IBM has announced the world's first sub-1 nanometer chip technology, a breakthrough in semiconductor manufacturing that promises major gains in computing power and energy efficiency. The new architect…
Qualcomm is acquiring Modular, the AI infrastructure startup founded by Chris Lattner and Tim Davis, for approximately $3.92 billion in stock. The deal, announced June 24, 2026, gives Qualcomm Modular…